A statistical test for comparing the degree of activation in the EEG due to sensory stimulation

A statistical test has been proposed to evaluate the significance in the degree of cortical activation due to sensory stimulation. Such test is based on a normalizing transformation to the coherence estimate between the stimulation signal and the EEG. Since it is not an exact transformation and also one of its parameter is unknown, a Monte Carlo simulation has been carried out to establish the most suitable way of employing the hypothesis test. The simulation indicates the minimum variance estimate of the unknown parameter as the more suitable one. The application of this methodology is illustrated with EEG signals (derivation O1) from 14 normal subjects during intermittent photic stimulation with flash at 6 Hz. In this case, the methodology proposed allows evaluating the degree of cortical activation, and so differences in the evoked responses due to the stimulation within subjects.